The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: firstname.lastname@example.org (Ge0)
Subject: Re: Thermal effects of potting compound?
Date: 22 Oct 2002 10:23:54 -0700
NNTP-Posting-Date: 22 Oct 2002 17:23:55 GMT
Perhaps I should have been more informative about what I was doing.
However, I always get assused of supplying too much detail. People
say I am too wordy. Anyway, here is the scoop:
I am using an IRL2905 D-PAK MOSFET to switch a pump motor on and off
in my assembly. The assembly will be potted AND placed in the engine
compartment of a vehicle. The circuit must remain reliable at
temperatures up to 125C.
The MOSFET is considered overkill for ambient temperature
applications, at my temperature extremes, and its surrounding
environment, I could be cutting it close using even this device.
The FET has a Rdson of 51mOHM at 125C. I'd like to have a lower Rdson
but it just isn't available at a 55V breakdown voltage AND with
reasonable availability. I am looking into using the IRLR3915 though.
It's Rdson is 29mOHM at 125C. The estimated Thermal resistance
junction to board is 50C/W given that the DPAK is mounted on FR4 board
with 2oz copper. The heat dissipating surface is roughly twice as
large as the solder pad for the tab of the device.
Under worse case conditions this device would be forced to dissipate
1.8W for approximately 500ms until the safety shutoff mechanism takes
place. Under worse case nominal operating conditions it would be
forced to dissipate 816mW for 8 seconds on, then 14 seconds off. This
duty cycle would repeat for 2 minutes 30 seconds and then the unit
would complete it cycle and shut off.
Thermal calculations show that junction temperature could climb to
165C given worse case nominal operating conditions. The thermal limit
of the device is 175C. This is too close for my comfort. Also, my
calculations were made assuming this sucker would be facing upright
and exposed to ambient air. The assembly it goes into will be potted
and sealed inside a plastic housing. It is an aftermarket device up
to the consumer to install. Who knows how it will be oriented within
the vehicle so i can't judge how the FET will be oriented in relation
to this. I made the assumption that being mounted upside down so heat
would have to transfer through the board as being worse case.
I'm sure I left a lot of holes so please ask more questions if you got
Thanks in advance,
"Tom Faloon" wrote in message news:...
> Hi George
> If you have a fan forcing sufficient cool air over both sides of the board,
> then turning the board upside down probably won't be a problem.
> If you don't have forced air cooling, then turning the board upside down
> will probably degrade the cooling performance considerably, because you will
> discourage convection. Some heat will be lost by convection off the upper
> side of the board, but the heat will have to pass through the PCB material,
> which is not a very good conductor. Cooling will be very inefficient
> compared with having the board 'right way up'.
> Whether 'upside down' cooling is good enough depends very much on your
> application - On factors such as, how much heat you are dissipating, case
> and PCB dimensions, case material and ventilation, and PCB layout. There is
> no simple formula. At the end of the day, you need to do some measurements!
> Re coating the board in 5mm of potting compound. This is sure to degrade
> cooling effectiveness. Talk to potting compound manufacturers, they may be
> able to suggest figures.
> You didn't say how much power you were dissipating in the D-PAK MOSFET. I
> would think twice before potting that, unless the power is very low! Apart
> from the effect on cooling, you will need a potting compound which can
> tolerate the surface temperature of the D PAK, and which won't crack, or
> disintegrate, with repeated heating & cooling.
> Tom Faloon
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup